Why QbD and Digitalization Are Foundations for Cell Therapies
An interview with Shin Kawamata of Japan’s Foundation for Biomedical Research and Innovation (FBRI) highlights exciting work to move cell therapy toward reliable and scalable commercialization.
Cell therapy involves modifying the human immune system by introducing cancer-specific T cells or regenerating damaged tissue by transplanting stem cells. Modified cells can be tailored to fight what have been considered until now to be untreatable illnesses. So called “personalized” treatments with cell therapy include taking cells from a patient, performing the genetic manipulation, and then putting them back into the same patient (autologous); another approach involves supplying modified cells to a pool of patients with a related makeup and disorder (allogeneic).
Shin Kawamata, Director at the Research and Development Center for Cell Therapy at FBRI in Kobe, Japan, is a researcher in the field of cell therapy and led the setup of the first commercial GMP facility in Japan. His facility gained regulatory approval and supplied the first chimeric antigen receptor T cells (CAR-T) product to young patients with leukemia. Kawamata graduated from Kyoto University (Division of Physics, Faculty of Science) and from Kobe University (School of Medicine). He started work as a clinician (hematology), then completed a pathology program at the Kyoto University Graduate School of Medicine and postdoctoral research at Stanford University School of Medicine. He was appointed as a Senior Researcher at the Institute of Biomedical Research and Innovation at the FBRI at Kobe.He has been an ISPE member since 2020.
Pharmaceutical Engineering® recently spoke with Kawamata about his work.
When did you first learn about cell therapy and what inspired you to get involved?
I am a hematologist, and working in the hospital 30 years ago, I was involved in bone marrow transplants for the treatment of leukemic patient using allogeneic stem cells. I saw how nature’s tools can be used to treat diseases and aid recovery beyond what was possible with operations and medication. During my postdoctorate in the US, I learned about gene modulation and translation for cells using a vector virus for the treatment of leukemic mouse. This was kind of a prototype of CAR-T, which I am today producing in a commercial facility, so there is a nice connection back to my earlier work and learnings in my career.
I see a beauty in working with the hematopoietic hierarchy. It is dynamic and often mysterious to understand the team work and collaboration of immunological cells. This area of medicine and biology charmed me very much and I became curious about it. Later in my career, I started to realize and apply myself to the many gaps that exist between the initial R&D of treatments and how to scale them in manufacturing to bring these products to patients.
What is your vision for the industry?
We need affordable treatment cost, accessibility for global patients, and the most effective and high quality products delivered to patients in the fastest possible time! As cell manufacturing is a live continuous process, it requires a different approach from conventional chemical compounds and small molecules manufacturing. I focus my work to improve the required methods, equipment, and systems for our emerging industry to scale to meet our vision. Recently with our partners, we launched a machine that was developed at FBRI to automate the entire manufacturing process for CAR-T, mesenchymal stem cells (MSC), and induced pluripotent stem cells (iPSC) products, the equipment reduced sterility risks, lowers facility costs and improves the speed and quality of treatments.
It’s been nearly 5 years since the first CAR-T therapy for cancer was approved by the US Food and Drug Administration. When you look at the cell and gene industry today, what’s most surprising to you?
In the last 5 years, there is suddenly a lot of competition and new entrants to cell therapy, but so far no one can dominate the market and everyone is still working on solving the basic quality and manufacturing issues and complications.
There has not been a consistent vision from the regulators and industry for immunotherapy and cell therapy. As an emerging industry, there is a lack of experts and links between the business leaders, providers of capital, and researchers. We need more educational support and collaboration with universities and organizations such as ISPE to prepare people for work in cell therapy. As a government funded organization, FBRI is charged with the advancement of knowledge using our research expertise and industrial operations background. As we have designed, commissioned, and run commercial CAR-T operations to supply products to patients in Japan, then we want to make our real-life and practical manufacturing experience available for use at global training bodies.
The other surprise is that cell manufacturing today relies on heavy manual operations, it reminds me of accounts from the 18th century textile mills as it is physically taxing and uncomfortable for the operators to perform in their full sterile gowning. Manual effort also has the associated high costs for setting up HVAC for the large grade B space required, which increases the investment for companies getting started. The other aspect is that there are few IT-based process control and documentation systems that can manage the dynamic nature of production and the mass of critical and timely data that must be assessed during manufacturing. I expected cell therapy to have an early adopter mindset that could “jump-ahead” rather than recreate and get burdened by legacy practices and inefficiencies. We need process automation and digitization equipment and tools at a cost and scale that is feasible for early stage therapeutic companies.
What has not changed in the cell therapy industry that frustrates you?
Cell manufacturing needs IT-based data formats, connections, and integrity. Currently our process operators spend a huge amount of effort filling out paper forms by hand, which wastes time and adds greater human error into our records. The quality and availability of data from paper is not sufficient for use in real-time process control and improvement, so we are missing opportunities to finetune our production for the running batch and to better orchestrate downstream activities. We are missing data in the required formats to apply advanced analytics on significant parameters and from QC data that would aid us to continuously optimize the product and process over time.
I feel that our industry needs to speed up, and I like to draw a comparison to the development of the semiconductor industry, where countries have been proactive and successful in setting up clusters of industry partners and contract manufacturing organizations (CMOs) that can make faster advancements by teaming up. I hope that commercial organizations, government, and regulators take a similarly strategic and long-term view to invest in these important treatments and build the human resource and manufacturing capabilities for the future.
Based on your experience of running commercial GMP phase operations for actual patients, what are the top 5 challenges that need to be addressed for cell manufacturing?
As our starting materials are not uniform, we must cope with their inherent variability, and this impacts everything from initial regulatory approval to the design of manufacturing systems and continuous control and verification approaches. We need practical GMP guidance for these products and we see that for our sector, quality by design (QbD) is mandatory to ensure quality through parameter measurement and adjustments– we are not able to examine cells destructively and so we need to examine the quality by sensing non-destructively or in process monitoring.
Manual processing and data availability have a high impact on the cost of sterility and operational reliability., Digitalization is thus vital for us to enable QbD and continuous monitoring. A huge amount of data is available, and we want to use this to identify and track the critical process parameters (CPPs) across both manufacturing and supply chain data.
With a relatively large inventory and number of consumables, we can have more than 50 items that all have to be tracked to the right production step and time, with their own shelf life and storage conditions. Many of these items must be imported, which increases reliance on supply chains and the management of suppliers.
There is a very high cost for IT systems deployment, including time and effort for implementing software. Affordable software is needed that can both comprehend the complex manufacturing process and easily integrate to finance, operations, and supply chain data to create assurance and value.
The challenges to train and retain people for working in a cell processing center (CPC) are very high. We need to expand our knowledge and improve learning and access to suitable courses, materials, tools and experience. Cell therapy requires a different approach and a good amount of hands-on work and organizations like FBRI can play a key bridging role as we promote learning and at the same run an operational facility.
You are involved in research to understand which cells show therapeutic effect after administration and to explore the parameters to assure cells possesses a therapeutic quality. How do you assure quality using a QbD-based approach?
Our aim for the control capability is to be able to predict and guarantee through data that cells being manufactured have the required therapeutic effect. We are approaching this using QbD theory to set up the CPPs through the key process stages and their acceptable variance in the design space (DS). As we must verify the validity of the CPPs and DS, this is done by conducting tests with existing manual-made cell products and animal models (for efficacy) for comparability study against cells manufactured through monitoring of CPPs with adaption of the process control limits within the DS. Once we confirm the quality by measurement, we aim to skip elaborative and destructive verification testing of the final products, or quality by testing (QbT). In our business, final product testing slows the shipping of products to patients who may be waiting for the treatment. The hope is that this mode of QA using real-time measurement and control can replace QbT. Not only do real-time control and feedback require digital systems and formats, we need fully automated manufacturing equipment that can perform the biomechanical adjustment of the process based on the data from the sensing and analytics system. So we very clearly see that to fully enable QbD, we need new technology in manufacturing plus the link to integrated real-time data. We believe we ultimately may generate new business and therapeutic benefits by connecting into big data systems such as the procurement of raw materials, supply chain and logistics optimization, and relating manufacturing conditions to ongoing clinical performance data.
Which advancements in sensor technologies do you consider as most important to enable QbD?
Liquid chromatography with mass spectrometry (LC-MS) is a key tool for us to perform the analysis of the cell metabolic system. Today this is done offline; we want to move online in the future or at least be able to sample and receive results in a fast and accurate fashion. Currently this involves tight coordination between manufacturing and quality control labs and systems. The other area we have a high interest is to apply machine learning (ML) and artificial intelligence (AI) for cell morphological image analysis taken by vision tools. We are excited at the use cases for ML/AI to aid the classification of cell morphology and measurement of mitochondrial activity to better understand the differentiation or aging of cells. One reason we are so focused on moving to fully electronic data across cell manufacturing is the increasing potential to analyze and find improvements that paper records simply cannot achieve. We are moving from thinking about individual sensor, measurement, and results and how to affect the process, to move to an open platform principle where data is always available and analysis can be done holistically across all data sources. In this way we can solve both the problems of system integration and the lack of use of data, for example, to find root cause when problems occur, as they occur, and find improvement opportunities far beyond what we understand by only looking at a single batch. Measurement technology is still the foundation for a data-centric approach and further advancements are required for continuous non-destructive sensors to monitor cell growth, cell morphology, lactose, cell type, and particular surface molecules. Manufacturing our products is like keeping fish in a tank: you have to monitor so many parameters, understand the CPPs, and constantly ensure that the fish all survive and look healthy.
You have been pioneering fully automated equipment for cell therapy manufacture. What is holding back the product development and market access for these technologies?
The base biological and mechanical functions are in place to allow for automation. Several teams and companies around the world are working on equipment for different types of products and specific phases of the overall production process such as cell expansion. I see the challenges to automate are no longer in the machinery required for manipulation of cell materials, but that across the industry we are still struggling to find a practical and effective approach to QbD), process analytical technology, and process validation: this is not cell therapy specific. There are examples and inspiration generally in the pharma industry for point use cases such as in biofermentation and continuous manufacturing. These show that sensor-equipment technology and the software control and analysis ability is there, but how to assemble this into an auditable solution is far from clear. From our experience, setting up the CPP is a key issue for product development. We are interested to progress this from the outset for a new product with a CMO/CDMO partner in order to show a proof of concept not only for the automation and software, but to include the entire approach to quality, verification, and how Industry 4.0 tools are central to pulling this all together.
Quality by Design is an old concept in pharma that the industry has struggled to adopt. How important is QbD for your vision of cell therapy?
QbD is indispensable for products that cannot undergo traditional process validation such as cell manufacturing. There are no standards to chemically test these against for our precursor starting materials. We have to ask ourselves how best to verify from collection from the patient, through the production process, and back to the patient that the product’s CQAs are valid. We believe the answer is in the accurate and timely collection of the right data and the premise of QbD that quality is assured by the right process, and that our process needs to continuously adapt. This is very different to working with small molecule APIs and ingredients, and so we need to wait for a while until cell therapy becomes popular and there is more impetus to change our thinking on topics such as validation and compliance.
Why do you see digitalization is such an important topics for cell therapy?
We think we have a lot of data in the industry; however, as that data is on paper, it is not useful. Today our expectations for what we can do with data in everyday life increases, yet at work in bioprocessing we are struggling with capturing records and forms that are impossible to evaluate for continuous monitoring and improvement. Cell therapy has a high need and urgency to make data from upstream and downstream processes such as clinical instantly available in and out of the CPC. This is simply not possible with paper, and current IT systems are not well designed for leveraging supply chain integration and to utilize the powerful analytical tools that are available from cloud providers.
Digitalization allows us to visualize and improve the manufacturing process and materials, link to the required inventory, warehouse control, change control, scheduling, staff mobilizing plans, and to optimize the operation cost in a real-time manner. Integrated systems also make it possible to connect to the data of raw material procurement and clinical data in the hospital that allows us to extract and pick up the QC parameters during manufacturing that are relevant to clinical efficacy. We are convinced that in cell therapy, the opportunities to understand the process through data will lead to new drug and process development. At the end of the day, digitalization promotes innovation and this is especially needed in the early stage of our industry.
What are the required capabilities for information technology for cell therapy that differ from the rest of the pharma and biopharma industry? What is missing in current software applications that are typically used?
Actually, we do not think the capabilities are really much different. But the need for them is higher in cell therapy. Process validation methods for conventional small molecule or chemical compounds manufacturing is not feasible due to the nature of live cell products. Instead, a QbD approach that requires higher level of digitalization such as data handling and integration of operation and QC data in IT format, ideally in cloud system is indispensable to faster cell manufacturing. But today, even in larger scale commercial manufacturing sites that have sufficient budgets, it should be noted that our experience with software is that it struggles with difficulties to integrate, high upfront and maintenance costs and demands a huge resource requirement to implement. Once in use we have seen a lack of flexibility in defining and executing our processes and so we are now researching new types of software that are event aware, able to use live process data and work proactively with our users. Our goal is that the Digitisation tools in cell therapy compliment and make operators more efficient, not slowing them down and making them frustrated with documentation.